Newcomers in plant communities: Interactions with soil and climate change

Soil infertility undermines the agriculture-based livelihoods in Malawi, where it is blamed for poor crop yields and the creation of cycles of poverty. Although technologies and management strategies have been developed to reverse the decline in soil fertility, they are under-used by smallholder farmers. This study was conducted to assess with farmers the performance of a range of maize-legume technologies and their benefits on soil fertility management in central Malawi. Farmer participatory experimentation was a focus of the study. The aim was to facilitate learning and the interpretation of experiences, improve the communication of information about the concepts and technologies to farmers, and provide insights for researchers.

 Using a combination of survey and participatory methods, 136 smallholder farmers from Chisepo were grouped into four resource groups, comprising of better-resourced (RG 1 with 6 farmers), medium resourced (RG 2, 14 farmers), less well-resourced (RG 3, 64 farmers) and least-resourced groups (RG 4, 52 farmers). Analysing their livelihoods for their effects on soil fertility revealed that soil fertility management is a complex activity which is influenced by ownership of assets. Farmers from RG 1 and RG 2 owned more resources including cattle, had larger fields, hired-in labour for timely farm operations, earned more income and invested far more in soil fertility improvement. Farmers from RG 3 and 4 (who are in the large majority) were resource constrained and did not invest adequately in improving soil fertility. They had large food deficits due to poor crop yields. Ganyu labour (casual work done for other farmers for food or cash) was their main strategy to reduce food deficits. Farmers from all the four RGs were interested in working with research to explore strategies to improve soil fertility. They tested various grain- and green-manure-legumes, and mineral N and P fertiliser on maize and the legumes for effects on crop productivity and soil fertility. Associated production risk and interest in technology adoption were assessed.

 On-farm evaluation was done on maize (cv. MH18) in rotation with pigeonpea cv. ICP 9145,intercropped with groundnut (cv. CG 7), (Mz/Pp+Gn); intercropped with tephrosia (Mz+Tv); intercropped with pigeonpea (Mz+Pp) and in rotation with mucuna (Mz/Mp). These technologies were compared with sole crop maize without fertiliser (Mz−Ft) or with 35 kg N ha^-1(Mz+Ft) in experiments with 32 farmers from the four RGs over four years. Economic and risk assessments were made. Maize grain yields (accumulated over the four years) were greater for farmers from RG 1 and 2 than RGs 3 and 4. Mz+Pp and Mz+Tv gave greater cumulative yields than Mz/Pp+Gn and Mz/Mp. The legumes improved maize grain yields by between 0.2 and 4 t ha^-1(P < 0.001) over Mz-Ft and additionally they gave legume grain to the household.Mz+Pp was less risky to all RGs, and applying 35 kg N ha^-1to the legumes resulted in Mz+Tv, Mz/Pp+Gn and Mz/Mp being least risky to RG 1, RG2 and RG 3. Farmers in RG 1 had the highest returns to labour (USconv2.info.8 day^-1with Mz-Ft and US.1 day^-1with Mz+Pp) and these increased to 1.9 and 1.7 respectively with 35 kg N ha^-1. Mz+Pp intercrop gave consistent positive returns across the RGs and was the only technology to provide positive returns to labour in RG 4. Use of pigeonpea was overall the least risky option, and was especially suited to least-resourced farmers.

 Application of phosphorus fertiliser (0, 20 kg P ha^-1) to legumes significantly (P = 0.05) increased grain and biomass yields for mucuna, groundnut, soyabean, Bambara groundnut and cowpea by 1.0, 0.8, 0.5, 1.0 and 0.3 t ha^-1compared with unfertilised plots. Cowpea and fertilised groundnut had larger yields in the home fields than middle fields, but other legumes performed better (P = 0.05) in the middle fields.

 Maize responses to small amounts of fertiliser (0, 15, and 30 kg N ha^-1and 0, 20 kg P ha^-1) in two weeding regimes showed that weeding twice significantly (P < 0.001) raised maize yields by 0.4 t ha^-1over weeding once (0.9 t ha^-1). Stover yields (significant at P < 0.001) were 2.3 and 1.6 t ha^-1respectively. Mean grain N kg ha^-1was 17.1 and 9.8 for plots weeded twice and once respectively while that of stover were 10.1 and 5.6 kg N ha^-1. Applying N at 15 kg N ha^-1increased maize yields, but the 30 kg N ha^-1increased yield only on more clay soils due to the effects of mid-season dry spells on sandy soils. Except for the physiological efficiency of N (PE_N), all agronomic indices of N use showed significant differences due to weeding (agronomic efficiency of applied fertiliser N (AE_N) at P < 0.001, recovery efficiency of applied N (RE_N) and partial factor productivity for N (PFP_N) at P < 0.01). The average PE_Nof 40.7and PFP_Nof 78.8 in plots weeded twice were within the ranges of 40–60 kg grain kg^-1N and 40–80 kg grain kg^-1N applied respectively. AE_Nand RE_N values of 38.7 and 0.9 respectively were above the common range of 10-30 kg grain kg^-1 N applied and 0.3-0.5 or 0.5–0.8 kg N kg^-1. Mean indices from plots weeded just once were all within the ranges stated above but lower than indices from plots weeded twice; suggesting the unsustainability of the use of fertiliser without means to raise its efficiency through better management or combination with organic resources. Weeding twice gave higher returns to labour (USconv2.info.30 day^-1) than weeding once (USconv2.info.05 day^-1) and gross margins of US5.00 and US.00 with labour taken into account respectively.Farmers need to ensure timely weeding to get decent efficiencies and returns from the fertiliser, especially in drier cropping seasons.

 Using surveys, focus group discussions and the analytical hierarchy process (AHP), adoption of the ten legumes introduced to farmers in Chisepo was assessed among 136 farmers in 2004 and 84 farmers in 2007. Thirty-five percent of the farmers in 2004 and 22% in 2007 had adopted at least one of the legumes, with food grain legumes predominantly soyabean, groundnut, pigeonpea and to a lesser extent Bambara groundnut and cowpea being most adopted. Mucuna and tephrosia were adopted by few farmers while sunnhemp and grahamiana were not adopted at all. Farmers from RGs 1 and 2 adopted more of the legumes than those from RG 3 and 4. Lack of consistent markets, a lack of seed for planting, as well as land and labour shortages were cited for weak adoption.

 Soil fertility management by smallholder farmers is influenced by ownership of assets and the majority poorer farmers fail to invest adequately in improving soil fertility. In the absence of such resources, grain legumes will play an important role as a source of both food and organic matter to improve soil fertility. The participatory methods used in the study helped farmers better understand some of the soil fertility concepts and options, including the legumes. There is need to focus on how to assist farmers with practical knowledge to help them best combine organic and mineral fertiliser resources for improving soil fertility, and to develop and promote new dual-purpose legume options that feed humans and the soil.

 Key words: Adoption, analytical hierarchy process, crop yield, financial returns, food security, household assets, legume integration, livelihoods, NP fertiliser, nitrogen use efficiency, production risk, resource groups, smallholder, soil fertility, weeding.